The Sex‐Stratified Effects of RSPO3 Expression on Body Fat and Blood Lipids: An Omics Mendelian Randomization Study

Yihua Guo; Zhihe Chen; Lanhui Huang; Zhongyun Zhang; Yujia Gong; Qian Yang; Jiqiu Wang; Jie Zheng

doi:10.1002/oby.70252

. 2026 Jul 8;34(8):1623–1634. doi: 10.1002/oby.70252

The Sex‐Stratified Effects of RSPO3 Expression on Body Fat and Blood Lipids: An Omics Mendelian Randomization Study

Yihua Guo ^1,², Zhihe Chen ^1,², Lanhui Huang ^1,², Zhongyun Zhang ^1,², Yujia Gong ^1,², Qian Yang ^1,^2,^3,^✉, Jiqiu Wang ^1,^2,^✉, Jie Zheng ^1,^2,^3,^✉

PMCID: PMC13422653 PMID: 42420700

ABSTRACT

Objective

This study examined sex‐stratified effects of R‐spondin 3 (RSPO3) expression levels on body fat and blood lipids and their potential mediation effects through sex hormones.

Methods

We performed a sex‐stratified genome‐wide association study (GWAS) of RSPO3 protein (N = 18,529 men and 21,323 women). We used two‐sample Mendelian randomization (MR) to examine associations of RSPO3 expression with the outcomes. Sex‐stratified MR estimates were compared using the pairwise z score test and Cochran's Q test. We conducted two‐step MR analyses to explore mediation effects through sex hormones.

Results

Sex‐stratified GWAS identified nine loci associated with RSPO3 protein levels, including three novel signals and four female‐stratified loci. In women, RSPO3 protein levels were associated with visceral adipose tissue (β = 0.22, 95% CI = 0.10 to 0.34), waist circumference (0.10, 0.05 to 0.15), hip circumference (0.10, 0.05 to 0.15), triglycerides (0.22, 0.18 to 0.26), high‐density lipoprotein cholesterol (−0.17, −0.21 to −0.13), and apolipoprotein A (−0.11, −0.15 to −0.06). We also observed sex differences in these associations. Two‐step MR showed that RSPO3 protein levels affected six outcomes via sex hormone‐binding globulin in women, with little evidence in men.

Conclusions

Our findings indicated that RSPO3 could be a drug target for regulating body fat distribution and blood lipids in women.

Keywords: blood lipids, body fat, Mendelian randomization, sex hormone‐binding globulin, sex‐stratified RSPO3 protein

1. Introduction

The manifestations of obesity have sex differences, including both body fat distribution and blood lipid profiles [1]. Previous studies have shown that canonical Wnt signaling factors, such as R‐spondin 1 (RSPO1), leucine rich repeat containing G protein‐coupled receptor 4 (LGR4), and low‐density lipoprotein receptor‐related protein 5 (LRP5), impact fat distribution and adiposity in a sex‐stratified manner [2]. R‐spondin 3 (RSPO3) is also a constituent of the R‐spondin family, which is integral to the modulation of Wnt signaling. This modulation influences various biological processes, including embryonic development and tissue regeneration [2]. A recent study investigated the role of RSPO3 in lipid metabolism, suggesting a potential connection to lipid‐related disorders such as hyperlipidemia and obesity [3]. Emerging evidence suggests that RSPO3 regulates fat distribution in a sex‐stratified manner in both humans and zebrafish [4]. However, this observed sex difference needs to be confirmed in large population‐based studies, and it remains unknown whether the effects of RSPO3 on lipids vary by sex.

Sex hormones have been implicated in driving sex differences in fat mass and distribution [5], regulating adipose tissue accumulation in a tissue‐specific manner [5]. Sex hormone‐binding globulin (SHBG), alternatively referred to as sex steroid‐binding protein or estradiol testosterone‐binding globulin, is mostly released by hepatocytes [5]. Recent research on SHBG has revealed its strong correlation with a range of metabolic traits, such as blood lipid levels, hepatic fat content, obesity, insulin resistance, and diabetes [6]. The effect of SHBG on fat distribution and blood lipids may exhibit sexual dimorphism, with post‐transcriptional mechanisms playing a key role in driving these sex differences [7]. Moreover, testosterone, the primary circulating androgen, significantly influences this process [7] and is a crucial hormone in the pathogenesis of metabolic diseases, including obesity [5]. In men, diminished testosterone levels are associated with decreases in lean muscle mass [5], insulin sensitivity, glucose tolerance, and high‐density lipoprotein cholesterol (HDL‐c) but increases in triglycerides (TG) and central adiposity [8]. In women, however, excessive androgen levels may contribute to the development of insulin resistance and metabolic dysfunction. These factors collectively elevate the risks of obesity and diabetes [9]. Thus, we hypothesized that sex hormones could mediate the sex‐stratified effects of RSPO3 on body fat and blood lipids.

Mendelian randomization (MR) studies employ genetic variants allocated at conception as instruments to infer causality, making them less susceptible to confounding and reverse causation than conventional observational studies [9]. When integrated with omics data, MR can provide strong genetic confirmation for potential causal genes, hence improving the success rate of drug trials [10].

The aim of this study is to investigate the sex‐stratified effects of genetically predicted RSPO3 protein levels and expression of RSPO3 on body fat and blood lipid phenotypes using MR. First, we conducted a sex‐stratified genome‐wide association study (GWAS) of circulating levels of RSPO3 protein to identify suitable genetic instruments for subsequent MR analyses. Second, we conducted two‐sample MR and a set of sensitivity analyses to explore the associations of interest. Finally, a two‐step MR was applied to elucidate whether genetically predicted RSPO3 protein levels and expression of RSPO3 affect body fat and blood lipid phenotypes via circulating levels of sex hormones.

2. Methods

2.1. Study Design

The study design is depicted in Figure 1. We initially performed sex‐stratified GWAS of circulating RSPO3 protein to identify suitable genetic instruments. Second, we performed sex‐stratified MR analyses to investigate the associations of genetically predicted RSPO3 protein levels and expression of RSPO3 with body fat and blood lipid phenotypes. Finally, a two‐step MR strategy was employed to examine whether the identified sex‐stratified effects on the outcomes were mediated by sex hormones.

Design of the current study. MR, Mendelian randomization; GWAS, genome‐wide association study; GTEx project, Genotype‐Tissue Expression project; eQTL, expression quantitative trait loci; pQTL, protein quantitative trait loci; SHBG, sex hormone‐binding globulin; TT, total testosterone; RSPO3, R‐spondin 3. [Color figure can be viewed at wileyonlinelibrary.com]

We followed the STROBE‐MR (Strengthening the Reporting of Observational Studies in Epidemiology Using Mendelian Randomization) guidelines to report our study [11]. Details are listed in Online Supplementary Information.

2.2. GWAS of Circulating RSPO3 Protein Levels

We conducted sex‐stratified GWAS in UK Biobank. Details of genotyping and quality control procedures are described in Online Supplementary Text, with a flowchart in Figure S1. Our sex‐stratified GWAS of circulating RSPO3 protein included 18,533 men and 21,323 women of European descent who are genetically unrelated. In UKB‐PPP, plasma protein profiling was measured using the Olink Explore 3072 Proximity Extension Assay (PEA) method [12], and protein measurements were expressed as Normalized Protein eXpression (NPX) values. We calculated associations of genetic variants on circulating RSPO3 protein using linear regression, adjusting for age, genotyping array, and the top 10 principal components. An additive model was employed using PLINK 1.9 [13]. Significant GWAS variants independently associated with RSPO3 protein levels were identified by excluding palindromic SNPs, applying a genome‐wide significance threshold (p < 5.0 × 10⁻⁸), and performing linkage disequilibrium (LD) clumping (r ² < 0.001 within a 1 Mb window).

Functional mapping and annotation of GWAS analyses illustrated SNP‐based and gene‐based Manhattan plots of sex‐stratified genome‐wide associations. To quantify the circulating level of RSPO3 protein attributed to common variations, we assessed heritability (h²) using LD score regression in R (ldscr, v.0.1.0) [14]. In addition, we utilized LD score regression to assess the genetic correlation (rg) between male‐ and female‐stratified RSPO3 protein and 33 outcomes (listed in ‘Genetic associations with outcomes’) along with three sex hormones (i.e., estradiol, SHBG, and total testosterone). The LD scores from the Pan‐UK Biobank dataset [15] were utilized as a reference. Specific steps included: (a) reformatting summary statistics; (b) filtering for the SNPs of Europeans in the Pan‐UK Biobank with associated alleles; (c) evaluating genetic associations. The LD scores from the Pan‐UK Biobank dataset were utilized as a reference.

2.3. Genetic Instruments for RSPO3 Protein Levels and Expression of RSPO3

We used expression quantitative trait loci (eQTL) and protein quantitative trait loci (pQTL) as instruments. The sex‐stratified eQTL data from three tissues (artery–tibial, adipose–subcutaneous, and cultured fibroblasts), acquired from the Genotype‐Tissue Expression (GTEx) project version 8 [16], comprised 838 people (281 women and 557 men) (Table S1). In GTEx, expression of the RSPO3 gene was measured using RNA‐seq with the Illumina TruSeq library construction protocol. The sex‐stratified pQTL (18,529 men and 21,323 women) were identified in the current study (Table S2). We mainly considered the following criteria to select eQTL and pQTL as instruments: (a) associated with the exposure of interest; (b) within the RSPO3 cis‐region (±1000 kb window); (c) using F statistics to assess the instrumental strength; (d) clumped at a LD r ² threshold of 0.001 to exclude variants with strong correlations to each other. Details of exposure‐specific criteria are listed in Online Supplementary Text. In addition, we evaluated heterogeneity by sex in each instrumental variable using the Cochran's Q test.

2.4. Genetic Associations With Mediators

We selected three sex hormones—estradiol, SHBG, and total testosterone—as candidate mediators. Genetic associations with these mediators were obtained from the largest sex‐stratified GWAS analyses publicly available at the time of analysis, with sample sizes ranging from 17,134 to 214,989 (Table S3), accessible via the IEU OpenGWAS project [17, 18]. We also considered sex hormones as exposures in mediation MR (described in Statistical analysis), and details of instrument selection are listed in Online Supplementary Text.

2.5. Genetic Associations With Outcomes

We chose two sets of phenotypes (illustrated in Figure S2). First, 26 body fat phenotypes included visceral adipose tissue (VAT), abdominal subcutaneous adipose tissue (ASAT), gluteofemoral adipose tissue (GFAT), fat mass and percentage, fat‐free mass from arm (left and right), leg (left and right), and trunk, basal metabolic rate, body fat percentage, whole body fat mass, fat‐free mass and water mass, body mass index (BMI), hip circumference (HC), and waist circumference (WC) (N ranging from 19,038 to 193,570). Second, for blood lipids, we included apolipoprotein A (apoA), apolipoprotein B (apoB), total cholesterol, HDL‐c, low‐density lipoprotein cholesterol (LDL‐c), lipoprotein A, and triglycerides (TG) (N ranging from 126,212 to 184,998). We obtained sex‐stratified GWAS summary statistics of 30 outcomes from the Neale Lab UK Biobank release [17, 18] (details in Table S3). In addition, GWAS summary statistics of VAT, ASAT, and GFAT were leveraged from the Cardiovascular Disease Knowledge Portal [19]. We also considered those phenotypes as exposures in reverse MR (described in Statistical analysis), and details of the instrument selection are listed in Online Supplementary Text.

2.6. Statistical Analysis

2.6.1. Two‐Sample MR Analysis

We conducted a two‐sample MR analysis to explore sex‐stratified causal associations of RSPO3 protein levels and expression of RSPO3 with body fat and blood lipids. For exposures having a single genetic instrument, the Wald ratio was used to estimate their effects on the outcomes [20]. To falsify MR assumptions, we conducted the following sensitivity analyses. First, we conducted the LD check to evaluate the possibility of confounding by LD structure [10]. An r ² of 0.7 between the eQTL or pQTL and any outcome variants was used as evidence of approximate colocalization [21]. Second, Debiased Inverse‐Variance Weighted (DIVW) [22], MR using Robust regression (MR‐Robust) [23], and MR‐Robust Adjusted Profile Score (MR‐RAPS) [24] were employed to rectify any possible breaches of the assumptions and to account for the influence of weak instrument bias. Third, we applied Steiger filtering to assess whether the exposure had the expected direction of effect, where the exposure influenced the outcome as a causal consequence [21].

To filter MR estimates with robust genetic evidence, we applied the false discovery rate (FDR) at α = 0.05 using the Benjamini–Hochberg method for multiple testing. MR estimates were different between men and women, as evidenced by a p value of pairwise z score < 0.05 [25] and Cochran's Q p value < 0.05.

2.6.2. Mediation MR

We conducted sex‐stratified two‐step MR [26] to evaluate whether the effects of genetically predicted RSPO3 protein levels and expression of RSPO3 on the tested body fat and blood lipids outcomes were mediated through circulating sex hormones. In step 1, we assessed the effect of both RSPO3 protein levels and expression of RSPO3 (exposure) on sex hormones (outcomes) using the Wald ratio (β1), given only one genetic variant was selected as the instrument (Figure S3A). We also applied the LD check to falsify the MR assumption regarding confounding by LD, with technical details described earlier. Those sex hormones showing MR evidence in step 1 were further considered in step 2 to estimate their effects on body fat and blood lipid phenotypes (β2) (Figure S3B). In step 2, we used inverse‐variance weighted (IVW) as the main analysis, and weighted median, weighted mode, and MR‐Egger regression as sensitivity analyses [27], given multiple genetic variants were used as instruments. The indirect effect (β1 × β2) was assessed using the product method, and the mediation proportion was calculated as the indirect effect divided by the total effect.

2.6.3. Reverse MR Analysis

To investigate whether body fat and blood lipid phenotypes have causal effects on RSPO3 protein levels, we conducted a reverse sex‐stratified MR analysis, where body fat and blood lipid phenotypes were treated as exposures, while circulating RSPO3 protein and sex hormones were the outcomes. IVW was used for the reverse MR analysis.

All MR analyses were conducted using Mendelian Randomization [28] and Two Sample MR [29] R packages.

3. Results

3.1. Sex‐Stratified GWAS Identified Variants Associated With RSPO3 Protein Levels

We discovered 10 variants robustly associated with RSPO3 protein levels (p < 5 × 10⁻⁸) (Figure 2A,B, Table S4). In the gene‐based association analysis, RSPO3 and PLG exhibited genetic signals in both sexes, while GP6, RDH13, SCGN, and SDC4 exhibited sex‐stratified associations exclusively in women (Figure 2C,D). Of note, variants within PLG, RDH13, and SCGN regions were not identified in the previous sex‐combined GWAS using UK Biobank [12].

Manhattan plot of sex‐stratified genome‐wide RSPO3 protein level associations. Functional mapping and annotation of sex‐stratified SNP‐based and gene‐based GWAS were conducted using the FUMA method (n = 18,533 in men, n = 21,323 in women). The x‐axis represents the genomic position, and the y‐axis represents the strength of association as represented by −log10 (p value). In panels A and B, the genome‐wide significance threshold was set at p < 5.0 × 10⁻⁸, whereas in panels C and D, p values were adjusted for multiple testing using false discovery rate (FDR) correction, with statistical significance defined as FDR q value < 0.05 and indicated by the red line. Female‐stratified candidate genes are denoted by blue text, and candidate genes found in both sexes are denoted by black text. (A) Sex‐stratified SNP‐based GWAS in men. (B) Sex‐stratified SNP‐based GWAS in women. (C) Sex‐stratified gene‐based GWAS in men. (D) Sex‐stratified gene‐based GWAS in women. [Color figure can be viewed at wileyonlinelibrary.com]

In terms of the fraction of RSPO3 protein levels attributed to common genetic variation, the SNP heritability was comparable in men (h² = 0.14, SE = 0.03, p = 2.03 × 10⁻⁶) and women (h² = 0.14, SE = 0.03, p = 2.98 × 10⁻⁷), alongside a robust positive genetic correlation between RSPO3 protein expression in both sexes (rg = 0.69, SE = 0.14, p = 1.01 × 10⁻⁶) (Table S5).

3.2. Causal Effects of Genetically Predicted RSPO3 Protein Levels and Expression of RSPO3 on Body Fat and Blood Lipids

MR analyses using sex‐stratified eQTL showed 7 and 15 robust MR signals (FDR < 0.05) in men and women, respectively, all with colocalization evidence (Figures 3A and 4A, Table S6). At least one MR sensitivity method showed marginal association results (p < 0.05) with consistent directionality for all phenotypes (Table S6). Using pairwise z score tests and Cochran's Q tests, we found that two RSPO3‐body‐fat pairs (HC and GFAT) and six RSPO3‐blood‐lipids pairs (including apoA, apoB, cholesterol, HDL‐c, LDL‐c, and TG) demonstrated distinguished MR estimates between men and women (Table S7).

Sex‐stratified Mendelian randomization (MR) results of the causal effects of genetically predicted RSPO3 expression and protein levels on body fat and blood lipid phenotypes. (A) Sex‐stratified two‐sample MR to estimate the causal effects of the expression of *RSPO3* on body fat and blood lipid phenotypes. The width of the connection line is expressed as the absolute value of Wald's ratio method β/p. (B) Sex‐stratified two‐sample MR to estimate the causal effects of RSPO3 protein levels on body fat and blood lipid phenotypes. Wald's ratio method was applied. WC, waist circumference; HC, hip circumference; TG, triglycerides; ApoA, apolipoprotein A; ApoB, apolipoprotein B; HDL‐c, high‐density lipoprotein cholesterol; LDL‐c, low‐density lipoprotein cholesterol; VAT, visceral adipose tissue; ASAT, abdominal subcutaneous adipose tissue; GFAT, gluteofemoral adipose tissue; RSPO3, R‐spondin 3. The asterisk indicates that the direction of the effect differs between sexes. [Color figure can be viewed at wileyonlinelibrary.com]

Summary of sex‐stratified Mendelian randomization (MR) results on the causal effects of genetically predicted RSPO3 expression and protein levels on body fat and blood lipid phenotypes. (A) Sex‐stratified two‐sample MR to estimate the causal effects of the expression of *RSPO3* on body fat and blood lipid phenotypes. (B) Sex‐stratified two‐sample MR to estimate the causal effects of RSPO3 protein levels on body fat and blood lipid phenotypes. The following lists the phenotypes that are significant for either men, women, or both: WC, waist circumference; HC, hip circumference; TG, triglycerides; Apo A, apolipoprotein A; Apo B, apolipoprotein B; HDL‐c, high‐density lipoprotein cholesterol; LDL‐c, low‐density lipoprotein cholesterol; VAT, visceral adipose tissue; GFAT, gluteofemoral adipose tissue; RSPO3, R‐spondin 3. [Color figure can be viewed at wileyonlinelibrary.com]

Using sex‐stratified pQTL as instruments, we identified nine robust MR signals (FDR < 0.05) with colocalization evidence from the LD check in women and three MR signals in men (Figures 3B and 4B, Table S8). Compared with the results based on eQTL, we discovered compelling evidence to substantiate the impact of elevated RSPO3 protein levels on increased VAT, WC, HC, apoB, cholesterol, LDL‐c, and TG and reduced apoA and HDL‐c in women (Figure 4, Table S9). Although no sex heterogeneity was observed in each instrumental variable, one RSPO3‐body‐fat (HC) pair and four RSPO3‐blood‐lipids pairs (apoB, HDL‐c, LDL‐c, TG) demonstrated non‐overlapping impact comparative estimates of men and women (pairwise z score p < 0.05 and Cochran's Q p < 0.05) both in eQTL and pQTL.

3.3. Mediation Effects of Sex Hormones

We identified SHBG as a mediator, and the detailed results of mediation MR are summarized here. In the first step, we explored the effects of genetically predicted RSPO3 protein levels and expression of RSPO3 on sex hormones in men and women separately. Using sex‐stratified RSPO3 eQTL as instruments, we showed that elevated genetically predicted expression of RSPO3 was associated with lower SHBG in women (LD check r ² = 0.699) but higher SHBG in men (Table S10). Using pQTL as the instrument to proxy the protein level of RSPO3, we validated the finding in women (Table S11).

In addition, genetically predicted expression of RSPO3 was associated with lower total testosterone in women, but not in men (Table S10). Using pQTL as an instrument, we observed no effects on testosterone in women and men nor evidence to support the influence of genetically predicted RSPO3 protein levels on estradiol (Table S11).

The pairwise z score tests and Cochran's Q tests showed that genetically predicted expression of RSPO3 exhibited sex‐stratified effects on SHBG and total testosterone between men and women (p < 0.05) (Table S12). Using pQTL as an instrument, the genetically predicted RSPO3 protein levels did not show sex‐stratified differences in SHBG and total testosterone (Table S13).

In the second step, we conducted sex‐stratified analysis to investigate the effects of sex hormones on body fat and blood lipid phenotypes. In women, SHBG was associated with 15 body fat phenotypes and 4 blood lipid phenotypes (p < 0.05). In men, SHBG was associated with 4 body fat phenotypes and 3 blood lipid phenotypes (p < 0.05) (Figure S4, Table S14). The pairwise z score tests and Cochran's Q tests suggested that SHBG had sex‐stratified effects on 25 body fat phenotypes and 7 blood lipid phenotypes (Table S15).

In addition, total testosterone was associated with 9 body fat phenotypes and 4 potential blood lipid phenotypes in women and with 15 body fat phenotypes and 3 potential blood lipid phenotypes in men (Figure S5, Table S16). The pairwise z score tests and Cochran's Q tests showed that total testosterone had sex‐stratified effects on 22 body fat phenotypes and 2 blood lipid phenotypes between men and women (pairwise z score p < 0.05 and Cochran's Q p < 0.5; Table S17). The association between estrogen and body fat phenotypes and blood lipid phenotypes was not pursued further, as the relevance assumption was not satisfied in the step 1 analysis.

MR sensitivity methods provided consistent findings in VAT, WC, HC, apoA, HDL‐c, and TG, with no evidence of horizontal pleiotropy (Table S14). VAT, WC, HC, apoA, HDL‐c, and TG demonstrated distinct impact estimates for men and women (pairwise z score p < 0.05 and Cochran's Q p < 0.05; Table S15).

In addition, we conducted a two‐step MR analysis to identify potential mediators. One mediator was identified. Increased genetically predicted RSPO3 protein levels were associated with lower SHBG in women, while lower SHBG led to an increase in VAT, WC, HC, and TG and a decrease in apoA and HDL‐c in women (Figure 5A). The proportions of mediation are presented in Table S18. However, in men, the mediation effect was much less conclusive.

RSPO3 regulates lipid and adipose metabolism through the Wnt signaling pathway, with SHBG serving as a mediator in this process. (A) Mediation results of this study. The left and right panels present the results at the gene and protein levels, respectively. The solid line represents the effects with strong Mendelian randomization evidence, and the dotted line represents the effects with weak Mendelian randomization evidence. (B) Possible mechanism. RSPO3 forms a complex by binding to LGR4 and LGR5, which in the left panel binds to the seven‐transmembrane Frizzled (FZD) and co‐receptor LRP5 to regulate Wnt signaling to affect body lipids and increase WC, HC, and VAT and in the right panel binds to the seven‐transmembrane FZD and co‐receptor LRP6 to regulate Wnt signaling to affect lipid levels and decrease ApoA and HDL‐c. SHBG can interact with Wnt signaling in vivo to regulate body lipid distribution and blood lipids. Wnt signaling regulates body fat distribution and lipids. WC, waist circumference; HC, hip circumference; Apo A, apolipoprotein A; HDL‐c, high‐density lipoprotein cholesterol; VAT, visceral adipose tissue; TG, triglycerides; SHBG, sex hormone‐binding globulin; RSPO3, R‐spondin 3. [Color figure can be viewed at wileyonlinelibrary.com]

3.4. Estimation of Reverse Effects of Body Fat and Blood Lipids on RSPO3 Protein Levels

All body fat and blood lipid phenotypes passed the Steiger filtering, as the instruments explained more variations in themselves than in RSPO3 protein levels (Tables S5 and S8). Reverse MR showed no evidence to support the effects of blood lipids on RSPO3 expression, except that lipoprotein A in men and women and eight body fat phenotypes in men and women were associated with RSPO3 protein levels (Table S19).

4. Discussion

Our sex‐stratified GWAS of RSPO3 protein confirmed pQTL in the RSPO3 region in both men and women, and further identified six novel loci that have not yet been reported in previous sex‐combined GWAS. Our MR study novelly found that genetically predicted RSPO3 protein levels and expression of RSPO3 appear to have putative female‐stratified unfavorable associations with body fat distribution (VAT, WC, HC, and TG) and blood lipid homeostasis (apoA and HDL‐c), with SHBG acting as their mediator.

Consistent with previous sex‐combined GWAS [12], RSPO3, GP6, and SDC4 loci achieved genome‐wide significance in either men or women in our sex‐stratified GWAS. We identified a novel locus PLG in our sex‐stratified GWAS of RSPO3 protein levels, which was not reported by the most recent sex‐stratified GWAS that used a very stringent threshold to identify sex differences [30]. After clumping the significant SNPs in the PLG region, we found one top variant rs9458016 in women and rs569706339 in men, with the directions of their effects being opposite. These results suggested that the sex‐stratified loci may remain obscured in sex‐combined GWAS. Genetically, the two SNPs (rs1936800 and rs1936801) in the cis region are most significantly associated with RSPO3 protein expression in males and females, respectively, and can serve as proxies for RSPO3 protein expression levels, exhibiting sex‐stratified effects. This approach has been widely used to evaluate causal associations of plasma proteins with diseases and their risk factors in omics MR [10]. Therefore, they can be utilized as instrumental variables for subsequent MR analyses of protein expression levels. The two SNPs are in perfect linkage disequilibrium (LD) with each other (LD r ² = 1 in 1000 Genomes Europeans). This implies that the sex‐differential effects of RSPO3 protein expression on obesity and lipid traits are mainly caused by different genetic effects of the same genetic signal.

Based on our instrument selection criteria in Online Supplementary Text, we identified three tissues—subcutaneous adipose, cell cultured fibroblasts, and artery tibial. Besides subcutaneous adipose, cell cultured fibroblasts mediate ventricular fibrosis, hypertrophy [31], and dysfunction in type 2 diabetes [32] and promote obesity‐induced adipose tissue fibrosis, while anterior tibial artery diameter can serve as an early marker of atherosclerosis [33].

Previous studies identified evidence supporting the sex‐stratified effects of RSPO3 on obesity. Consistent with previous studies [34], our findings confirmed associations of RSPO3 with higher WC and HC, but lower BMI, body fat percentage, and leg fat mass. Notably, the effects on WC and HC were more pronounced in women, indicative of sexual dimorphism. Genetically predicted RSPO3 protein levels and expression of RSPO3 affect HC in opposite directions between sexes, suggesting sex‐stratified biological interactions. Our study was less consistent with the Adipose Tissue Knowledge Portal meta‐analyses of 21 male‐stratified and 5 female‐stratified datasets, revealing that RSPO3 was associated with obesity in men but not in women [34]. We, however, demonstrated significantly stronger effects on obesity modulation in women than men, and such an inconsistency could be due to insufficient female‐specific data in the previous study. We novelly found that genetically predicted expression of RSPO3 influenced visceral fat solely in women, with no corresponding effect in men. One possible explanation could be that women are known to possess a greater percentage of subcutaneous fat and a lesser percentage of visceral fat in comparison to men [35]. This relatively smaller amount of visceral fat in women may have enhanced the detection of a genetic signal. Another possible explanation is that the abundance of this pathway differs between subcutaneous and visceral fat. For example, LGR4 is more highly expressed in visceral fat than in subcutaneous fat [36], whereas CTNNB1 shows stronger expression in subcutaneous fat compared with visceral fat [37]. However, whether this expression pattern exhibits sex‐specific differences remains unclear.

We observed contrasting effects of genetically predicted RSPO3 protein levels and RSPO3 expression on arm fat‐free mass and HC, showing a positive association in women but a negative association in men. Notably, this pattern is consistent with established sex‐specific fat distribution trends, whereby males preferentially accumulate fat in the upper body, whereas females predominantly store fat in the lower body [38]. Furthermore, expression of RSPO3 in subcutaneous adipose tissue was found to be higher in men than in women. In females, RSPO3 activates Wnt/β‐catenin signaling, increasing the sensitivity of gluteal adipocytes to apoptotic stimuli, thereby inhibiting the expansion of the gluteal and thigh fat depot. This effect was not observed in males [4]. These findings suggest that RSPO3 plays a differential role in the regulation of fat distribution. We did not identify previous studies exploring the effects of genetically predicted RSPO3 protein levels and expression of RSPO3 on lipids, though genetic variations in the RSPO3 gene correlated with abdominal obesity were linked to dyslipidemia [39]. Our study suggested that genetically predicted RSPO3 protein levels and expression of RSPO3 could indicate unfavorable effects on blood lipids, particularly in women. These sex‐stratified differences may relate to sexually dimorphic expression of RSPO3, which is higher in men, yet women show a stronger physiological response to its expression [4].

Our observed associations can be explained by the following mechanisms (illustrated in Figure 5B). For body fat, LRP5 expression plays an important role, which is higher in abdominal progenitor cells compared to gluteal progenitor cells. It acts by titrating β‐catenin signal strength to modulate adipose progenitor biology in a depot‐specific manner, thereby promoting lower body fat accumulation [40]. RSPO3 forms a complex by binding to LGR4 and LGR5 [41], which interacts with the seven‐transmembrane Frizzled (FZD) and the co‐receptor LRP5 to regulate Wnt signaling [40], affecting body fat distribution and increasing WC, HC, and VAT [42]. This may be due to the different expression patterns of RSPO3 and LRP5 in WAT, with higher expression in visceral fat compared to subcutaneous fat [40]. For blood lipids, mutations in LRP6 are associated with dyslipidemia, including high LDL‐c and elevated TG [42]. The RSPO3 complex binds to the seven‐transmembrane FZD and the co‐receptor LRP6 [40], regulating Wnt signaling by antagonizing Kremen/DKK‐dependent LRP6 internalization [43], thereby influencing blood lipid levels, specifically increasing TG and decreasing apoA and HDL‐c [44].

Regarding the observed mediators, both male and female individuals with obesity tend to exhibit lower levels of SHBG, which contributes to the dysregulation of sex hormones and metabolic disturbances, including insulin resistance and altered lipid profiles [45]. Our findings are consistent with previous studies, demonstrating that higher SHBG levels are associated with increased BMI, WC, HC, and HDL‐c, but lower body fat mass and TG [46, 47]. Our study novelly found that increased SHBG was linked to decreased WC and HC in women, but the opposite effects were observed in men. Therefore, the effects of SHBG extend from being a carrier for sex hormones to metabolic regulation [47]. Biologically, SHBG can interact with Wnt signaling in the body to regulate body fat distribution and blood lipids [48] (illustrated in Figure 5B).

Our findings have an important clinical implication. Currently, no obesity medications target RSPO3. The only medication targeting RSPO3 is rosmantuzumab, a monoclonal antibody that has completed a phase 1 trial in cancers (NCT02482441) [49] (Table S20). Our study suggested that RSPO3 could be a potential drug target for body fat and blood lipids management, requiring further studies to evaluate its efficacy and safety.

The strengths of this study included cross‐validation of results using both eQTL and pQTL data and multiple MR sensitivity analyses. Furthermore, a two‐step MR approach was applied to explore potential sex‐stratified mediators underlying these effects.

However, limitations do exist. First, the F statistic of the sex‐stratified eQTL of RSPO3 is lower than 10, suggesting potential bias due to weak instruments [50]. However, we applied robust MR methods for weak instruments to mitigate such bias. Second, we used a limited number of genetic variants to proxy RSPO3 protein levels and the expression of RSPO3, which may constrain the application of MR sensitivity analyses and reduce statistical power. Third, we used a larger sample size of pQTL data in women compared to men, and this imbalance in sample size may affect the statistical power of our MR analysis [25]. Fourth, considering that we possess eQTL data just from European ancestry solely, it is imperative to obtain data from other ancestries to evaluate the generalizability of ancestry‐specific effects on body fat and blood lipids [25]. Fifth, our two‐sample MR analyses based on pQTL data of the UK Biobank could be biased due to sample overlap, as some sex‐stratified GWAS analyses of outcomes were also from the UK Biobank. However, the percentage of sample overlap was approximately 0.1, which may not have a meaningful impact on the MR estimates. Finally, we considered three mediators that may not capture the full spectrum of potential mediatory hormones. Therefore, although we identified SHBG as a plausible mediator, other unexamined hormones could also play a role in our observed associations.

5. Conclusion

Our findings highlighted sexually dimorphic effects of genetically predicted RSPO3 protein levels and expression of RSPO3 on adiposity and lipid metabolism, potentially mediated by SHBG, suggesting RSPO3 as a potential drug target for managing body fat and blood lipids.

Author Contributions

J.Z. and J.W. contributed to the study's concept and design. Y.G. and J.Z. performed the MR analyses, validated, and visualized the results. Y.G., Q.Y., and J.Z. wrote the first draft of the manuscript. Z.C. independently validated the results based on summary‐level data. L.H. contributed to genotyping and quality control of circulating RSPO3 protein GWAS. J.Z., J.W., and Y.G. revised the manuscript with critical revisions from all authors. Z.Z. and Y.J.G. assisted in revising the manuscript. Q.Y., J.W., and J.Z. contributed to the funding acquisition and supervision. J.Z. is the guarantor of this study.

Funding

This work was supported by grants from the National Key Research and Development Program of China (2021YFA1301103), the Noncommunicable Chronic Diseases‐National Science and Technology Major Project (2024ZD0531500, 2024ZD0531502, 2023ZD0508302), and Lingang laboratory (LGL‐2616‐01). The National Key Research and Development Program of China (2022YFC2505200, 2022YFC2505203) sponsored this work. Q.Y. and J.Z. are supported by the National Natural Science Foundation of China (32500519, 32570728).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Quality control procedures of UK Biobank.

Figure S2: Outcomes of human body mapping.

Figure S3: Research flowchart.

Figure S4: Sex‐stratified Mendelian randomization results of the causal effects of sex hormone‐binding globulin on body fat and blood lipid phenotypes.

Figure S5: Sex‐stratified Mendelian randomization results of the causal effects of total testosterone on body fat and blood lipid phenotypes.

OBY-34-1623-s001.docx^{(1.1MB, docx)}

Table S1: Genetic instruments of RSPO3 eQTL.

Table S2: Genetic instruments of RSPO3 pQTL.

Table S3: Source of obesity indicators and sex hormones genome‑wide association study data.

Table S4: Significant GWAS loci for RSPO3 protein levels (p < 5 × 10−8).

Table S5: Genetic correlation between RSPO3 protein levels in males and females.

Table S6: Effects of genetically predicted expression of RSPO3 on body fat and blood lipid phenotypes in sex‐stratified datasets with four methods.

Table S7: Pairwise z score tests and Cochran's Q tests of genetically predicted expression of RSPO3 with body fat and blood lipid phenotypes Mendelian randomization estimates in two sexes.

Table S8: Effects of genetically predicted RSPO3 protein levels on body fat and blood lipid phenotypes in sex‐stratified datasets.

Table S9: Pairwise z score tests and Cochran's Q tests of genetically predicted RSPO3 protein levels with body fat and blood lipid phenotypes Mendelian randomization estimates in two sexes.

Table S10: Effects of genetically predicted expression of RSPO3 on sex hormones in sex‐stratified datasets.

Table S11: Effects of genetically predicted RSPO3 protein levels on sex hormones in sex‐stratified datasets.

Table S12: Pairwise z score tests and Cochran's Q tests of genetically predicted expression of RSPO3 with sex hormones Mendelian randomization estimates in two sexes.

Table S13: Pairwise z score tests and Cochran's Q tests of genetically predicted RSPO3 protein levels with sex hormones Mendelian randomization estimates in two sexes.

Table S14: Effects of SHBG on body fat and blood lipid phenotypes in sex‐stratified datasets with five methods.

Table S15: Pairwise z score tests and Cochran's Q tests of SHBG with body fat and blood lipid phenotypes Mendelian randomization estimates in two sexes.

Table S16: Effects of total testosterone on body fat and blood lipid phenotypes in sex‐stratified datasets with five methods.

Table S17: Pairwise z score tests and Cochran's Q tests of total testosterone with obesity phenotypes Mendelian randomization estimates in two sexes.

Table S18: Mediation Mendelian randomization analysis outcomes.

Table S19: Reverse Mendelian randomization results.

Table S20: Drug target validation of RSPO3.

OBY-34-1623-s002.docx^{(523.4KB, docx)}

Data S1: Supporting Information.

OBY-34-1623-s003.docx^{(43.1KB, docx)}

Acknowledgments

This study was conducted using the UK Biobank Resource under application number 104086. We thank all the participants for their involvement in UK Biobank.

Contributor Information

Qian Yang, Email: yq13248@rjh.com.cn.

Jiqiu Wang, Email: wangjq@shsmu.edu.cn.

Jie Zheng, Email: zj12477@rjh.com.cn.

Data Availability Statement

GWAS summary statistics used in this study were leveraged from publicly available studies except for sex‐stratified GWAS of RSPO3 protein levels in UKB‐PPP. Details of the data source were listed in Table S3. Full information on how to access UK Biobank data can be found at its website (https://www.ukbiobank.ac.uk/use‐our‐data/).

References

1. Regensteiner J. G. and Reusch J. E. B., “Sex Differences in Cardiovascular Consequences of Hypertension, Obesity, and Diabetes,” Journal of the American College of Cardiology 79 (2022): 1492–1505. [DOI] [PMC free article] [PubMed] [Google Scholar]
2. Sun Y., Zhang J., Hong J., et al., “Human RSPO1 Mutation Represses Beige Adipocyte Thermogenesis and Contributes to Diet‐Induced Adiposity,” Advanced Science 10 (2023): 2207152. [DOI] [PMC free article] [PubMed] [Google Scholar]
3. Kim K.‐A., Wagle M., Tran K., et al., “R‐Spondin Family Members Regulate the Wnt Pathway by a Common Mechanism,” Molecular Biology of the Cell 19 (2008): 2588–2596. [DOI] [PMC free article] [PubMed] [Google Scholar]
4. Loh N. Y., Minchin J. E. N., Pinnick K. E., et al., “RSPO3 Impacts Body Fat Distribution and Regulates Adipose Cell Biology In Vitro,” Nature Communications 11 (2020): 2797. [DOI] [PMC free article] [PubMed] [Google Scholar]
5. Bourebaba N., Ngo T., Śmieszek A., Bourebaba L., and Marycz K., “Sex Hormone Binding Globulin as a Potential Drug Candidate for Liver‐Related Metabolic Disorders Treatment,” Biomedicine & Pharmacotherapy 153 (2022): 113261. [DOI] [PubMed] [Google Scholar]
6. Yang J., Zhou J., Liu H., et al., “Blood Lipid Levels Mediating the Effects of Sex Hormone‐Binding Globulin on Coronary Heart Disease: Mendelian Randomization and Mediation Analysis,” Scientific Reports 14 (2024): 11993, 10.1038/s41598-024-62695-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
7. Basaria S., “Male Hypogonadism,” Lancet 383 (2014): 1250–1263. [DOI] [PubMed] [Google Scholar]
8. Kelly D. M. and Jones T. H., “Testosterone: A Metabolic Hormone in Health and Disease,” Journal of Endocrinology 217 (2013): R25–R45. [DOI] [PubMed] [Google Scholar]
9. Sebo Z. L. and Rodeheffer M. S., “Testosterone Metabolites Differentially Regulate Obesogenesis and Fat Distribution,” Molecular Metabolism 44 (2020): 101141. [DOI] [PMC free article] [PubMed] [Google Scholar]
10. Zheng J., Haberland V., Baird D., et al., “Phenome‐Wide Mendelian Randomization Mapping the Influence of the Plasma Proteome on Complex Diseases,” Nature Genetics 52 (2020): 1122–1131. [DOI] [PMC free article] [PubMed] [Google Scholar]
11. Skrivankova V. W., Richmond R. C., Woolf B. A. R., et al., “Strengthening the Reporting of Observational Studies in Epidemiology Using Mendelian Randomization: The STROBE‐MR Statement,” JAMA 326 (2021): 1614–1621. [DOI] [PubMed] [Google Scholar]
12. Sun B. B., Chiou J., Traylor M., et al., “Plasma Proteomic Associations With Genetics and Health in the UK Biobank,” Nature 622 (2023): 329–338. [DOI] [PMC free article] [PubMed] [Google Scholar]
13. Purcell S., Neale B., Todd‐Brown K., et al., “PLINK: A Tool Set for Whole‐Genome Association and Population‐Based Linkage Analyses,” American Journal of Human Genetics 81 (2007): 559–575. [DOI] [PMC free article] [PubMed] [Google Scholar]
14. Bulik‐Sullivan B., Finucane H. K., Anttila V., et al., “An Atlas of Genetic Correlations Across Human Diseases and Traits,” Nature Genetics 47 (2015): 1236–1241. [DOI] [PMC free article] [PubMed] [Google Scholar]
15. Karczewski K. J., Gupta R., Kanai M., et al., “Pan‐UK Biobank Genome‐Wide Association Analyses Enhance Discovery and Resolution of Ancestry‐Enriched Effects,” Nature Genetics 57 (2025): 2408–2417. [DOI] [PMC free article] [PubMed] [Google Scholar]
16. Oliva M., Muñoz‐Aguirre M., Kim‐Hellmuth S., et al., “The Impact of Sex on Gene Expression Across Human Tissues,” Science 369 (2020): eaba3066. [DOI] [PMC free article] [PubMed] [Google Scholar]
17. Elsworth B., Lyon M., Alexander T., et al., “The MRC IEU OpenGWAS Data Infrastructure,” BioRxiv (2020): 2020.08.10.244293, 10.1101/2020.08.10.244293. [DOI] [Google Scholar]
18. Hemani G., Zheng J., Elsworth B., et al., “The MR‐Base Platform Supports Systematic Causal Inference Across the Human Phenome,” eLife 7 (2018): e34408. [DOI] [PMC free article] [PubMed] [Google Scholar]
19. Agrawal S., Wang M., Klarqvist M. D. R., et al., “Inherited Basis of Visceral, Abdominal Subcutaneous and Gluteofemoral Fat Depots,” Nature Communications 13 (2022): 3771. [DOI] [PMC free article] [PubMed] [Google Scholar]
20. Burgess S., Small D. S., and Thompson S. G., “A Review of Instrumental Variable Estimators for Mendelian Randomization,” Statistical Methods in Medical Research 26 (2017): 2333–2355. [DOI] [PMC free article] [PubMed] [Google Scholar]
21. Xu S., Wang P., Fung W. K., and Liu Z., “A Novel Penalized Inverse‐Variance Weighted Estimator for Mendelian Randomization With Applications to COVID‐19 Outcomes,” Biometrics 79 (2023): 2184–2195. [DOI] [PMC free article] [PubMed] [Google Scholar]
22. Zhang D., Gao B., Feng Q., et al., “Proteome‐Wide Association Studies for Blood Lipids and Comparison With Transcriptome‐Wide Association Studies,” HGG Advances 6 (2025): 100383. [DOI] [PMC free article] [PubMed] [Google Scholar]
23. Rees J. M. B., Wood A. M., Dudbridge F., and Burgess S., “Robust Methods in Mendelian Randomization via Penalization of Heterogeneous Causal Estimates,” PLoS One 14 (2019): e0222362. [DOI] [PMC free article] [PubMed] [Google Scholar]
24. Verbanck M., Chen C.‐Y., Neale B., and Do R., “Detection of Widespread Horizontal Pleiotropy in Causal Relationships Inferred From Mendelian Randomization Between Complex Traits and Diseases,” Nature Genetics 50 (2018): 693–698. [DOI] [PMC free article] [PubMed] [Google Scholar]
25. Yang Q., Yang Q., Wu X., et al., “Sex‐Stratified Genome‐Wide Association and Transcriptome‐Wide Mendelian Randomization Studies Reveal Drug Targets of Heart Failure,” Cell Reports Medicine 5 (2024): 101382. [DOI] [PMC free article] [PubMed] [Google Scholar]
26. Carter A. R., Sanderson E., Hammerton G., et al., “Mendelian Randomisation for Mediation Analysis: Current Methods and Challenges for Implementation,” European Journal of Epidemiology 36 (2021): 465–478. [DOI] [PMC free article] [PubMed] [Google Scholar]
27. Sanderson E., Glymour M. M., Holmes M. V., et al., “Mendelian Randomization,” Nature Reviews Methods Primers 2 (2022): 6. [DOI] [PMC free article] [PubMed] [Google Scholar]
28. Yavorska O. O. and Burgess S., “MendelianRandomization: An R Package for Performing Mendelian Randomization Analyses Using Summarized Data,” International Journal of Epidemiology 46 (2017): 1734–1739. [DOI] [PMC free article] [PubMed] [Google Scholar]
29. Bowden J., Davey Smith G., and Burgess S., “Mendelian Randomization With Invalid Instruments: Effect Estimation and Bias Detection Through Egger Regression,” International Journal of Epidemiology 44 (2015): 512–525. [DOI] [PMC free article] [PubMed] [Google Scholar]
30. Koprulu M., Wheeler E., Kerrison N. D., et al., “Sex Differences in the Genetic Regulation of the Human Plasma Proteome,” Nature Communications 16 (2025): 4001. [DOI] [PMC free article] [PubMed] [Google Scholar]
31. Tuleta I., Hanna A., Humeres C., et al., “Fibroblast‐Specific TGF‐β Signaling Mediates Cardiac Dysfunction, Fibrosis, and Hypertrophy in Obese Diabetic Mice,” Cardiovascular Research 120 (2024): 2047–2063. [DOI] [PMC free article] [PubMed] [Google Scholar]
32. Kohda H., Tanaka M., Shichino S., et al., “Novel Cell‐To‐Cell Communications Between Macrophages and Fibroblasts Regulate Obesity‐Induced Adipose Tissue Fibrosis,” Diabetes 74 (2025): 1135–1152. [DOI] [PMC free article] [PubMed] [Google Scholar]
33. Abd El Aziz R., Fawzy M. W., Khalil N., Abdel Atty S., and Sabra Z., “Vascular Affection in Relation to Oxidative DNA Damage in Metabolic Syndrome,” Advances in Endocrinology and Metabolism 9 (2018): 43–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
34. Zhong J., Zareifi D., Weinbrenner S., et al., “adiposetissue.org: A Knowledge Portal Integrating Clinical and Experimental Data From Human Adipose Tissue,” Cell Metabolism 37 (2025): 566–569. [DOI] [PubMed] [Google Scholar]
35. Chen Z., Yu H., Shi X., et al., “Functional Screening of Candidate Causal Genes for Insulin Resistance in Human Preadipocytes and Adipocytes,” Circulation Research 126 (2020): 330–346. [DOI] [PMC free article] [PubMed] [Google Scholar]
36. Zhang Z., Wang L., Qiao H., et al., “Cryo‐EM Structure of the Full‐Length LGR4‐RSPOs Complex and a Targeting Nanobody for Anti‐Obesity Therapy,” Nature Communications 16 (2025): 8406. [DOI] [PMC free article] [PubMed] [Google Scholar]
37. Chen M., Lu P., Ma Q., et al., “CTNNB1/β‐Catenin Dysfunction Contributes to Adiposity by Regulating the Cross‐Talk of Mature Adipocytes and Preadipocytes,” Science Advances 6 (2020): eaax9605. [DOI] [PMC free article] [PubMed] [Google Scholar]
38. Karastergiou K., Smith S. R., Greenberg A. S., and Fried S. K., “Sex Differences in Human Adipose Tissues – The Biology of Pear Shape,” Biology of Sex Differences 3 (2012): 13. [DOI] [PMC free article] [PubMed] [Google Scholar]
39. Schleinitz D., Böttcher Y., Blüher M., and Kovacs P., “The Genetics of Fat Distribution,” Diabetologia 57 (2014): 1276–1286. [DOI] [PubMed] [Google Scholar]
40. Kriz V. and Korinek V., “Wnt, RSPO and Hippo Signalling in the Intestine and Intestinal Stem Cells,” Genes 9 (2018): 20. [DOI] [PMC free article] [PubMed] [Google Scholar]
41. Carmon K. S., Gong X., Lin Q., Thomas A., and Liu Q., “R‐Spondins Function as Ligands of the Orphan Receptors LGR4 and LGR5 to Regulate Wnt/β‐Catenin Signaling,” Proceedings of the National Academy of Sciences of the United States of America 108 (2011): 11452–11457. [DOI] [PMC free article] [PubMed] [Google Scholar]
42. Loh N. Y., Neville M. J., Marinou K., et al., “LRP5 Regulates Human Body Fat Distribution by Modulating Adipose Progenitor Biology in a Dose‐ and Depot‐Specific Fashion,” Cell Metabolism 21 (2015): 262–272. [DOI] [PMC free article] [PubMed] [Google Scholar]
43. Binnerts M. E., Kim K.‐A., Bright J. M., et al., “R‐Spondin1 Regulates Wnt Signaling by Inhibiting Internalization of LRP6,” Proceedings of the National Academy of Sciences of the United States of America 104 (2007): 14700–14705. [DOI] [PMC free article] [PubMed] [Google Scholar]
44. Mani A., Radhakrishnan J., Wang H., et al., “LRP6 Mutation in a Family With Early Coronary Disease and Metabolic Risk Factors,” Science 315 (2007): 1278–1282. [DOI] [PMC free article] [PubMed] [Google Scholar]
45. Hammond G. L., “Plasma Steroid‐Binding Proteins: Primary Gatekeepers of Steroid Hormone Action,” Journal of Endocrinology 230 (2016): R13–R25. [DOI] [PMC free article] [PubMed] [Google Scholar]
46. Garcés C., de Oya I., Lasunción M. A., López‐Simón L., Cano B., and de Oya M., “Sex Hormone–Binding Globulin and Lipid Profile in Pubertal Children,” Metabolism 59 (2010): 166–171. [DOI] [PubMed] [Google Scholar]
47. Azene Z. N., Davis S. R., McNeil J. J., Tonkin A. M., Handelsman D. J., and Islam R. M., “Estrone, Sex Hormone Binding Globulin and Lipid Profiles in Older Women: An Observational Study,” Climacteric 26 (2023): 114–120. [DOI] [PMC free article] [PubMed] [Google Scholar]
48. Elbers J. M. H., de Jong S., Teerlink T., Asscheman H., Seidell J. C., and Gooren L. J. G., “Changes in Fat Cell Size and In Vitro Lipolytic Activity of Abdominal and Gluteal Adipocytes After a One‐Year Cross—Sex Hormone Administration in Transsexuals,” Metabolism 48 (1999): 1371–1377. [DOI] [PubMed] [Google Scholar]
49. Fischer M. M., Yeung V. P., Cattaruzza F., et al., “RSPO3 Antagonism Inhibits Growth and Tumorigenicity in Colorectal Tumors Harboring Common Wnt Pathway Mutations,” Scientific Reports 7 (2017): 15270. [DOI] [PMC free article] [PubMed] [Google Scholar]
50. Lawlor D. A., Harbord R. M., Sterne J. A. C., Timpson N., and Davey Smith G., “Mendelian Randomization: Using Genes as Instruments for Making Causal Inferences in Epidemiology,” Statistics in Medicine 27 (2008): 1133–1163. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials